Garry Tan on Y Combinator Startup Podcast
How to Build the Future: Demis Hassabis
A technically serious, Demis-focused interview that takes Garry into continual learning, reasoning, and AI-driven scientific discovery—areas largely absent from his other appearances. Worth hearing for the guest discussion, but skippable for fans seeking substantial new material about Garry himself.
Some of it is like how much of it will be autonomous versus I mean I don't think we'd see autonomous first. We would actually probably see people in this room operating at a thousand X and then that's what you should see first. And then many of you, you know, there'll be like games companies or you know, other types of companies that have built some kind of best selling app, best selling game using these tools. That's what you should see first. And then more of that will get automated. I mean some of it is like there's a human in there and then the human doesn't want to say that the agents did it yet.
Garry Tan interviews Demis Hassabis about the missing capabilities required for AGI, including continual learning, memory, reasoning, and agents. The conversation also covers efficient and open models, AI-enabled scientific discovery, and the kinds of deep-tech companies that may remain defensible as foundation models improve.